Markyt evaluation system supports the evaluation of automatic entity and relationship predictions, typically within the scope of a community text mining challenge. Prediction evaluation is based on conventional quality metrics (i.e. precision, recall and f-score) and prediction matches and mismatches can be interactively explored.

Now you can evaluate your predictions on demand against the test set of the The BioCreative V.5

BioCreative V

The BioCreative V.5 Challenge will address:

  1. The development and evaluation of named entity recognition systems for chemical and gene/protein entity mentions and
  2. The proof-of concept implementation of a continuous text annotation metaservice, called BeCalm.
(BARR) track

The proposed Biomedical Abbreviation Recognition and Resolution (BARR) track - 2017 has the aim to promote the development and evaluation of biomedical abbreviation identification systems by providing Gold Standard training, development and test corpora manually annotated by domain experts with abbreviation-definition pairs within abstracts of biomedical documents written in Spanish.

BARR track asks participating teams to test processing text systems that are able to detect explicit occurrences of abbreviation-definition pairs. Please, click in the next button in order to access to the corpora and BARR resources:

BioCreative VI

BioCreative VI: Text mining chemical-protein interactions (CHEMPROT) A range of different types chemical-protein/gene interactions are of key relevance for biology, including metabolic relations (e.g. substrates, products) inhibition, binding or induction associations. Our aim is to promote research in this field, and to focus on chemical-protein interactions that might be of relevance for precision medicine as well as for drug discovery and basic biomedical research. This task will consist of two subtasks:

  1. Chemical-protein interaction pair detection task: Extracting relations between chemical entities and protein/genes belonging to at least one of a pre-defined set of relation types.
  2. Chemical-protein interaction type detection task: Providing for previously detected interaction pairs (of task 1) the corresponding relation type qualifier).
Markyt evaluation system multipropose

Cloud evaluation

Evaluate automatic predictions everywhere 24/7 using only a browser.

Markyt evaluation system  servers

Prediction evaluation

Markyt is able to evaluate automatic entity and relation predictions.

Markyt evaluation system  Evaluation

Interactive evaluation

Users may exhaustively explore prediction results.